Kaynes Technology — fat margins, thin returns, and a very expensive bet on becoming a chipmaker

The Pulse

Kaynes is the high-end cousin of India’s contract-manufacturing boom — instead of assembling cheap phones, it builds complex, mission-critical electronics for defence, aerospace, space, railways, medical and industrial customers, which is why it earns a fat ~16% operating margin where commodity manufacturers earn 4%. The paradox is that those fat margins produce thin returns: return on capital is only ~13% and return on equity under 10%, because the business is hungry for both fixed capital and working capital, and it is now pouring enormous sums into two audacious new bets — semiconductor packaging (an OSAT plant at Sanand) and high-end circuit boards. FY26 revenue grew 33% to ₹3,626 crore on a ₹9,000-crore order book, but free cash flow was negative ₹1,800 crore, management has missed and then abandoned its revenue guidance, and a research house publicly questioned its accounting — prompting a special rebuttal call. The stock has duly crashed ~60% from its peak. The whole case is a bet that today’s low-return, cash-guzzling investment phase gives way to a high-return, IP-rich electronics champion. It might; it is far from proven.

The Business

Most of what Kaynes does is “integrated electronics manufacturing services” (IEMS) — but the word that matters is integrated. Where a commodity manufacturer just assembles to a customer’s spec, Kaynes does conceptual design, process engineering, manufacturing, and lifecycle support, and it does it for hard, regulated, low-volume-high-value end markets: defence and aerospace electronics, railway signalling (including the Kavach train-collision-avoidance system), satellites, nuclear, medical devices, industrial and automotive. That mix is the source of its edge. These are markets with real qualification barriers — a defence or railway customer cannot casually switch its electronics supplier — and they reward design capability over sheer scale, which is why Kaynes earns four times the margin of a phone assembler. The order book tells the same story: about ₹9,000 crore, non-cancellable, with no single customer more than ~6% — genuinely diversified, growing roughly 50% a year.

What makes the company distinctive — and risky — is its decision to climb much further up the value chain into capital-intensive, IP-owning businesses. There are four new bets. The biggest is OSAT (outsourced semiconductor assembly and test) — a ₹3,200-crore chip-packaging plant at Sanand, heavily subsidised (50% from the central government, 20% from the state), which has already shipped India’s first commercial multichip module and signed up names like Alpha & Omega Semiconductor, Infineon and L&T Semiconductor. The second is high-density (HDI) circuit boards — a ₹3,700-crore, government-approved plant making the ~76-layer boards that defence and aerospace need, shipping from FY27 with PLI support. The third is smart metering, and the fourth is AI-based railway safety. The strategic logic is coherent: own more of the electronics chain, capture higher-margin IP, and ride India’s import-substitution and “make-in-India semiconductor” push. The promoter, Ramesh Kunhikannan, holds ~53.5% (down about 10 points over three years), and the founding family runs it alongside MD Dr. Muthukumar and CFO Jairam Sampath.

How Management Thinks

The leadership is ambitious and visionary — the recurring frame is a deliberate transformation from “EMS-led” to a “vertically integrated, innovation-led, design-first” company that owns intellectual property rather than just billing conversion charges. That ambition is real and, in the mission-critical niches, credible. But the period under review exposed genuine cracks in execution and communication that temper the read on management quality. They guided FY26 revenue to ₹4,500 crore, cut it to ₹4,000 crore, and delivered ₹3,626 crore — then dropped absolute guidance altogether, committing only to “twice market growth,” to visible analyst frustration (one bluntly said “the call is all over the place”). They badly misjudged the smart-metering business, whose AMISP model pays only on physical installation and so blew a hole in working capital when installations stalled — a mistake they are now reversing by pivoting to a device-and-software model.

The most serious episode was an accounting-scrutiny event: a research note (from Kotak) questioned the goodwill and intangibles from the Iskraemeco metering acquisition, the working-capital metrics, and the cash-flow reconciliation, prompting management to convene a dedicated rebuttal call. To their credit, CFO Jairam Sampath defended the accounting line by line and conceded one genuine error (a missed related-party disclosure in a subsidiary); to the watchful eye, the fact that the question arose at all — alongside a disputed working-capital number (management’s 87 days versus an analyst’s 157), debtor days of 154, and contingent liabilities of ₹5,200 crore — is a yellow flag for a company already consuming cash at this rate. The honest moment came when Sampath admitted on free cash flow: “I’m not sure in the near-term… we’ll have to reinvest for growth.” That is the truthful summary of the whole company. Capital allocation is all-in on the transformation: every rupee of profit (no dividend), plus subsidies, bridge debt and likely equity, is going into the new plants. The strategy is defensible; the execution and disclosure have wobbled.

Where It’s Going

The trajectory is a steep, deliberate investment J-curve. The core IEMS business should keep compounding on its large order book (management’s standing ambition is $1 billion of revenue by FY28 and $2 billion by FY30, though after the guidance misses these are best treated as aspirations rather than commitments). The real swing factor is whether the new bets convert. The OSAT chip-packaging plant is the headline: subsidy-de-risked, first modules shipped, marquee customers signed, with management talking of meaningful revenue from FY27 and over ₹2,500 crore of five-year visibility — but it is export-first, early-stage, and semiconductor packaging is a genuinely hard business to scale. The HDI board plant adds a higher-margin defence/aerospace leg from FY27. If these ramp at the margins management implies, returns on capital should rise from today’s depressed levels and the cash burn should reverse.

The tensions are exactly the mirror image of the opportunity. The returns are low now (ROCE ~13%, ROE under 10%) and the company is burning cash hard, so the entire investment case is deferred to an unproven future ramp. Working capital remains stretched and is the single most-watched number — management has promised positive operating cash flow and lower working-capital days by FY27 without sacrificing margin, a promise its recent credibility makes worth verifying rather than trusting. Layer on the guidance misses, the accounting episode, the foreign-investor exodus (FIIs down from 14% to 7% while retail doubled), and a still-rich valuation, and the picture is of a genuinely interesting, well-positioned business whose execution and capital discipline have not yet earned the benefit of the doubt.

The Four Checks

1. Quality & moat (gate) — 6/10. A genuinely better business than commodity EMS. The mission-critical niche (defence, aerospace, space, railways, medical) brings real qualification barriers, design-led switching costs, and pricing power — evidenced by 16% margins and a diversified, non-cancellable ₹9,000-crore order book. The new OSAT and HDI bets, if they work, deepen the moat into genuinely scarce capabilities. Held to a 6 because it is still fundamentally a contract manufacturer (the customer owns end-demand), the high-IP transformation is mostly ahead of it, and recent execution wobbles cloud the picture.

2. Returns on incremental capital & runway — 5/10. The weak spot, and the heart of the bear case. Despite fat margins, returns on capital are only ~13% and ROE under 10%, because the business devours fixed and working capital — free cash flow was minus ₹1,800 crore in FY26. The runway is genuinely long and expanding (semiconductors, PCBs, defence, import substitution), but a rupee reinvested today earns a modest, and partly unproven, return. The whole thesis rests on those returns rising as the new plants ramp — a J-curve that has not yet turned.

3. Capital allocation for the stage — 5/10. Mixed. The strategic direction — climb the value chain into subsidy-backed semiconductors and high-end boards, fund it by retaining all earnings — is ambitious and arguably right for the stage. But the execution undercuts it: a smart-metering model that blew up working capital (now being reversed), repeated guidance misses, an accounting-scrutiny episode requiring a public rebuttal, and deeply negative free cash flow. Genuine ambition, real governance and discipline question marks.

4. Price — 3/10. Demanding, even after the fall. At ₹3,077 the stock trades at ~56x earnings and ~4.3x book on a sub-10% ROE business that is burning cash — a multiple that only makes sense if the semiconductor/PCB transformation delivers a large step-up in returns. The ~60% crash from the ₹7,705 peak has removed much of the froth, but in absolute terms you are still paying a high-growth-compounder price for a business whose returns are currently low and whose execution is unproven.

Engine score: 16/30 (moat 6 + reinvestment 5 + allocation 5). Price 3.

Sources

• Concalls read: Q2 FY26 (call 5 Nov 2025), a special business-update/accounting-rebuttal call (8 Dec 2025), Q3 FY26 (6 Feb 2026), Q4/FY26 (May 2026) — cleaned BSE transcripts, the backbone of this digest (order book, segment detail, the OSAT/PCB/metering bets, working capital, guidance, and the accounting controversy). Note the Dec-2025 call was a non-standard event convened to rebut a research note.
• Annual reports: FY23, FY24, FY25 — all three extracts were heavily trimmed (chairman/MD letters and MD&A survived mostly as headers); the usable signal was the strategy framing (the EMS→design-led/IP shift), the FY23 revenue-mix baseline (~92% OEM contract manufacturing), the FY25 capex disclosure, and customer-concentration notes. The qualitative read leans on the concalls.
• Snapshot: screener.in (consolidated, logged-out) fetched 2026-06-13 09:26 IST.
• Gaps flagged: trimmed ARs; a disputed working-capital figure (management 87 days vs an analyst’s 157 — noted in-text); the contested “core growth ex-metering” math the CFO agreed to reconcile; significant other income (₹154 cr, likely treasury) flatters profit; no dividend; CFO absent (health) from the May 2026 call; logged-out snapshot. Promoter Kunhikannan family ~53.5%.
• Research dumps: vault/Sources/Earnings/Kaynes Technology India Ltd/.